Currently, software that is customizable is difficult if not impossible to test in all of its potential configurations. In many cases, one reason for this is that much of the code running as part of the system is authored by developers from different organizations tailoring the product in ways unforeseeable to the developers writing the customizable core system and to other developers customizing the system. Currently, there is no good way for a customizable system, especially a large-scale implementation of such a system, to efficiently identify incompatible customizations and provide corresponding solutions to heal the system so as to keep functionality in line with expectations.
The discussion above is merely provided for general background information and is not intended for use as an aid in determining the scope of the claimed subject matter.